The development of antimicrobial agents has led to a significant decrease in morbidity and mortality from infectious diseases in this century. This important public health contribution has been largely due to the widespread use of antibiotics that target specific nutrient, cell wall, DNA, RNA and protein biosynthetic pathways that are peculiar to pathogenic bacteria. However, in recent years the capacity to manage infectious diseases has been threatened by the emergence of bacterial strains that are no longer susceptible to currently available antimicrobial agents (see Files, 1999, Chest. 115:3S-8S). Maintenance of the public heath mandates that new antimicrobial agents need to be developed to counter these emerging resistant bacteria in order for effective infectious disease management procedures to remain in place.
A heterogeneous group of host-derived antimicrobial peptides have drawn attention as possible new therapeutic agents (see Hancock, R. E., 1999, Drugs 57:469-473). These peptides play an important role in innate vertebrate immunity against infection. For example, cationic antimicrobial peptides constitute as much as 18% by weight of total neutrophil protein. They are also found in high concentrations on damaged mucosal surfaces. In general these host-derived cationic peptides fit into one of four structural categories: (i) xcex2-sheet structures that are stabilized by multiple disulfide bonds (e.g., human defensin-1), (ii) covalently stabilized loop structures (e.g., bactenecin), (iii) tryptophan (Trp)-rich, extended helical peptides (e.g., indolicidin), and (iv) amphipathic xcex1-helices (e.g., the magainins and cecropins) (see Hwang and Vogel, 1998, Biochemistry and Cell Biology 76:235-246). Recently a new class of antimicrobial peptides, the cathelicidins, that utilize all of these structural motifs and are clearly important in host defense against infection has been described (Ganz and Lehrer, 1997, Current Opinion in Hematology 4:53-58).
The cathelicidins are a remarkably diverse collection of molecules that derive from prepropeptides sharing a highly conserved N-terminal propeptide segment that have been described in humans, cattle, sheep, rabbits, mice, and pigs (see Hwang and Vogel, 1998, Biochemistry and Cell Biology 76:235-246). The conserved propeptide segment of approximately 100 amino acids shares sequence similarity with the porcine protein cathelin, a putative cysteine protease inhibitor, hence the family name. The C-terminal domain encodes an antimicrobial peptide motif similar to one of those described above, depending upon the host and tissue that it is associated with. Cathelicidins are stored in neutrophil granules as propeptides (lacking antimicrobial activity in this form), with neutrophil activation leading to elastase-mediated endoproteolytic cleavage and generation of the C-terminal antimicrobial peptide. The human cathelicidin, referred to alternatively as FALL-39, hCAP18, LL-37, or CAMP, in its processed (active) form is a 37-amino acid amphiphilic xcex1-helical cationic peptide (see Zanetti, Gennaro and Romeo, 1995, FEBS Letters 374:1-5). Expression of LL-37 has been detected in human neutrophils, testicular cells, respiratory epithelia, and in keratinocytes at sites of inflammation.
The amphipathic cationic peptides of the xcex1-helical class demonstrate minimal bactericidal concentrations (MBCs) in the xcexcg/mL range (levels equivalent to other antimicrobial agents) and are able to kill a broad range of gram-negative and gram-positive bacterial pathogens, including those that are highly resistant to multiple antibiotics (see Hancock, R. E., 1999, Drugs 57:469-473). The mechanism by which these peptides kill bacteria proceeds in a two step process by first binding to the negatively charged bacterial surface and driving these bound peptides into the bacterial membrane, thereby disrupting its structural integrity. For gram-negative organisms, cationic antimicrobial peptides have the added advantage of binding lipopolysaccharide (LPS), thereby detoxifying its endotoxic activity (see Scott, Yan, and Hancock, 1999, Infection and Immunity 67:2005-2009). The hallmark of amphipathic cationic xcex1-helical antimicrobial peptides is their capacity to fold into an amphipathic secondary structure that presents a hydrophilic face with a net positive charge of at least +2. A number of different amino acid sequence combinations allow a peptide to achieve this characteristic structure. Consequently, hundreds of host-derived amphipathic cationic xcex1-helical peptides have been described to date all showing limited sequence homology at the level of primary sequence comparison (see Hwang and Vogel, 1998, Biochemistry and Cell Biology 76:235-246).
In contrast to host derived antimicrobial peptides, which have evolved with the express purpose of killing bacteria, a novel class of antimicrobial peptides derived from discrete segments of the lentiviral transmembrane (TM) protein cytoplasmic tail has been described that have not evolved for the same purpose as host-derived peptides (see Beary et al., 1998, Journal of Peptide Research 51:75-79; Comardelle et al., 1997, AIDS Research and Human Retroviruses 13:1525-1532; Miller et al., 1993, AIDS Research and Human Retroviruses 9:1057-1066; Miller et al., 1993, Virology 196:89-1000; Tencza et al., 1995, Virology 69:5199-5202; Tencza et al., 1997, Antimicrobial Agents and Chemotherapy 41:2394-2398; Tencza et al., 1997, AIDS Research and Human Retroviruses 13:263-269; Yuan et al., 1995, Biochemistry 34:10690-10696). These peptides are referred to as lentiviral lytic peptides (LLPs) with the prototypical LLP being LLP1 (amino acids 828-856 of the HIV-1 viral isolate HXB2R Env). LLP1 is derived from the 28-residues encoded by the C-terminal portion of the HIV-1 TM protein that, when modeled as an xcex1-helix, demonstrates amphipathic character with clearly delineated cationic and hydrophobic faces. Among the many antimicrobial peptides currently described in the literature, LLP1 is most homologous chemically to the magainins and the human cathelicidin, LL37.
LLP1 has been studied for its calmodulin-binding and antibacterial properties. LLP1 binds to host cell Ca2+-saturated calmodulin with near nanomolar affinity and this property has been correlated with the inhibition of T-cell activation, suggesting that these peptides may dampen an inflammatory response (see Beary et al., 1998, Journal of Peptide Research 51:75-79; Miller et al., 1993, AIDS Research and Human Retroviruses 9:1057-1066; Tencza et al., 1995, Virology 69:5199-5202; Tencza et al., 1997, AIDS Research and Human Retroviruses 13:263-269; Yuan et al., 1995, Biochemistry 34:10690-10696). LLP1 antibacterial activity has been investigated by surveying diverse gram-negative and -positive bacterial isolates. This analysis demonstrates that LLP1 has antibacterial activity which is equal to, or more potent than magainin-2. These isolates included methicillin and vancomycin resistant strains as well as other strains that were highly resistant to multiple antibiotics (see Tencza et al., 1997, Antimicrobial Agents and Chemotherapy 41:2394-2398). The lysis of bacteria by LLP1 is rapid, nearly sterilizing a suspension of 1xc3x97105 colony-forming units of Pseudomonas aeruginosa or Staphylococcus aureus within 60 seconds of exposure (see Tencza et al., 1997, Antimicrobial Agents and Chemotherapy 41:2394-2398). The mechanism of LLP1 action is thought to perturb negatively charged bacterial membranes, and to a lesser extent, neutral mammalian cell membranes. The predilection of the peptide for bacterial cells over mammalian cell membranes forms the basis for its selective toxicity.
Single amino acid changes in the LLP1 profoundly affect its calmodulin binding and antibacterial activity (see Tencza et al., 1995, Virology 69:5199-5202; Tencza et al., 1999, Journal of Antimicrobial Chemotherapy 44:33-41). In general, amino acid substitutions in the parent LLP1 sequence of basic residues to acidic residues decrease both calmodulin binding and bactericidal activities. Similarly, altering single hydrophobic residues to hydrophilic residues also decreased both of these activities. Furthermore, dimerization through disulfide bond formation of a single Cys found within the LLP1 parent sequence significantly increased its activity for S. aureus (see Tencza et al., 1999, Journal of Antimicrobial Chemotherapy 44:33-41). Finally, decreasing the length of the LLP1 dimer to 21 residues (peptide bis-TL1) reduced its red blood cell lysis activity without significantly reducing its antibacterial activity (see Tencza et al., 1999, Journal of Antimicrobial Chemotherapy 44:33-41). These data suggest that the LLP1 parent sequence can be engineered for increased potency and selectivity. The potential for this engineering forms the basis for this invention.
The present invention is directed to peptides having antimicrobial activity (xe2x80x9cantimicrobial peptidesxe2x80x9d). In one embodiment of the invention three antimicrobial peptides which are derived from, and are analogs of, the LLP1 peptide parent sequence corresponding to amino acids 828-856 of the HIV-1 viral isolate HXB2R Env have been described and include SA-5 (SEQ ID NO:1), LSA-5 (SEQ ID NO:2) and WLSA-5 (SEQ ID NO:3) (see Table 1 below). The antimicrobial activity of other LLP1 peptide analogues has been previously described (see Tencza et al, 1999, Journal of Antimicrobial Chemotherapy 44:33-41, U.S. Pat. No. 5,714,577 of Montelaro et al. and U.S. Pat. No. 5,945,507 of Montelaro et al.).
In another embodiment of the invention, the antimicrobial peptides are LLP1 analogs having modifications based on the following principles: (i) optimizing amphipathicity, (ii) substituting arginine (Arg) on the charged face and/or valine (Val) or tryptophan (Trp) on the hydrophobic face with another amino acid, and (iii) increasing peptide length (referred to collectively herein as (LBU) peptides, (e.g. LBU-1, SEQ ID NO:4; LBU-2, SEQ ID NO:5; LBU-3, SEQ ID NO:6; LBU-3.5, SEQ ID NO:7; LBU-4, SEQ ID NO:8, WLBU-1, SEQ ID NO:9, WLBU-2, SEQ ID NO:10; WLBU-3, SEQ ID NO:11; and WLBU-4, SEQ ID NO:12; see Table 1). The LBU peptides deviate greatly from the parent LLP1, for example, LBU-2 and LBU-3 deviate from the parent LLP1 sequence by greater than 90%.
The LLP1 analogue peptides and the LBU peptides (collectively referred to herein as xe2x80x9cengineered LLPsxe2x80x9d (eLLPs)) of the present invention have a broader spectrum of activity (i.e., the ability to kill highly resistant bacteria) and increased potency (i.e., lowering the molar concentration required to kill bacteria) when compared with previously described LLP1 analogs. The eLLPs of the present invention are highly inhibitory to microorganisms under physiologic salt concentrations, function in the presence of synovial fluid, and demonstrate only minimal toxicity in animal models. As a result, the eLLPs may be defined as selective antimicrobial agents. In addition, the peptides of the present invention function by disrupting bacterial membranes and are active when bound to a solid phase. The ability of these peptides to maintain activity when bound to a solid phase is a significant advantage over conventional antibiotics in that these peptides may be useful as coatings on sterile devices such as prostheses or catheters where it would be advantageous to prevent bacterial biofilm nucleation.